Comparison of asymptotic variances of inhomogeneous Markov chains with application to Markov chain Monte Carlo methods

DOI10.1214/14-AOS1209zbMATH Open1319.60152arXiv1307.3719MaRDI QIDQ464192FDOQ464192

Randal Douc, Florian Maire, Jimmy Olsson

Publication date: 17 October 2014

Published in: The Annals of Statistics (Search for Journal in Brave)

Abstract: In this paper, we study the asymptotic variance of sample path averages for inhomogeneous Markov chains that evolve alternatingly according to two different $pi$-reversible Markov transition kernels $P$ and $Q$. More specifically, our main result allows us to compare directly the asymptotic variances of two inhomogeneous Markov chains associated with different kernels $P_i$ and $Q_i$, $iin0,1$, as soon as the kernels of each pair $(P_0,P_1)$ and $(Q_0,Q_1)$ can be ordered in the sense of lag-one autocovariance. As an important application, we use this result for comparing different data-augmentation-type Metropolis-Hastings algorithms. In particular, we compare some pseudo-marginal algorithms and propose a novel exact algorithm, referred to as the random refreshment algorithm, which is more efficient, in terms of asymptotic variance, than the Grouped Independence Metropolis-Hastings algorithm and has a computational complexity that does not exceed that of the Monte Carlo Within Metropolis algorithm.

Full work available at URL: https://arxiv.org/abs/1307.3719

Recommendations

• Asymptotic Variance and Convergence Rates of Nearly-Periodic Markov Chain Monte Carlo Algorithms
• Minimising MCMC variance via diffusion limits, with an application to simulated tempering
• Kernel estimators of asymptotic variance for adaptive Markov chain Monte Carlo
• Optimal variance reduction for Markov chain Monte Carlo
• Examples comparing importance sampling and the Metropolis algorithm

zbMATH Keywords

asymptotic varianceMonte Carlo methodsinhomogeneous Markov chainsPeskun orderingpseudo-marginal algorithms

Mathematics Subject Classification ID

Computational methods in Markov chains (60J22) Monte Carlo methods (65C05) Numerical analysis or methods applied to Markov chains (65C40) Analysis of variance and covariance (ANOVA) (62J10) Markov chains (discrete-time Markov processes on discrete state spaces) (60J10)

Cites Work

• The Calculation of Posterior Distributions by Data Augmentation
• Title not available (Why is that?)
• The pseudo-marginal approach for efficient Monte Carlo computations
• Title not available (Why is that?)
• Monte Carlo sampling methods using Markov chains and their applications
• A note on Metropolis-Hastings kernels for general state spaces
• General state space Markov chains and MCMC algorithms
• On variance conditions for Markov chain CLTs
• Yet Another Look at Harris’ Ergodic Theorem for Markov Chains
• Ordering and improving the performance of Monte Carlo Markov chains.
• Optimum Monte-Carlo sampling using Markov chains
• Convergence properties of pseudo-marginal Markov chain Monte Carlo algorithms
• The Multiple-Try Method and Local Optimization in Metropolis Sampling

Cited In (13)

• Variational formulas for asymptotic variance of general discrete-time Markov chains
• Computable bounds of exponential moments of simultaneous hitting time for two time-inhomogeneous atomic Markov chains
• A hybrid scan Gibbs sampler for Bayesian models with latent variables
• Quantitative Convergence Rates for Subgeometric Markov Chains
• Analysis of two-component Gibbs samplers using the theory of two projections
• On the use of Markov chain Monte Carlo methods for the sampling of mixture models: a statistical perspective
• Peskun-Tierney ordering for Markovian Monte Carlo: beyond the reversible scenario
• Stability of noisy Metropolis-Hastings
• Asymptotic Variance and Convergence Rates of Nearly-Periodic Markov Chain Monte Carlo Algorithms
• Variational principles for asymptotic variance of general Markov processes
• Bayesian learning of weakly structural Markov graph laws using sequential Monte Carlo methods
• Comparison of Markov chains via weak Poincaré inequalities with application to pseudo-marginal MCMC
• Introduction to ``On the use of Markov chain Monte Carlo methods for the sampling of mixture models by R. Douc, F. Maire, J. Olsson

This page was built for publication: Comparison of asymptotic variances of inhomogeneous Markov chains with application to Markov chain Monte Carlo methods

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q464192)